CMF-Impute: an accurate imputation tool for single-cell RNA-seq data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) enables the study of cellular heterogeneity.
• scRNA-seq data analysis is challenged by technical noise, particularly dropout events (excessive zero counts).

Purpose of the Study:

• To propose a novel method, CMF-Impute, for imputing dropout entries in scRNA-seq data.
• To evaluate CMF-Impute's performance against existing state-of-the-art methods.

Main Methods:

• Developed CMF-Impute, a collaborative matrix factorization-based method.
• Tested CMF-Impute on six real and three simulated scRNA-seq datasets.
• Compared CMF-Impute with five other methods using metrics like SSE, Pearson correlation, ARI, and NMI.

Main Results:

• CMF-Impute accurately imputes dropout entries in simulated datasets.
• Achieved superior cell classification accuracy on real datasets compared to other methods.
• Demonstrated effectiveness in reconstructing cell-cell and gene-gene correlations.
• Showed power in inferring cell lineage trajectories.

Conclusions:

• CMF-Impute is a robust method for handling dropout events in scRNA-seq data.
• The method enhances downstream analyses such as cell classification and trajectory inference.
• CMF-Impute offers improved biological insights from scRNA-seq data.